INDEX

Explanations

a specific string related to verifying humanity, often found in a captcha-like context asking users to confirm they are not robots

oai_token-act-pair · gpt-3.5-turbo

instances of the article "a" in various contexts

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.10.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 unexpectedly

-0.71

 unequ

-0.69

 innumerable

-0.66

almost

-0.64

Effects

-0.64

scenes

-0.64

amorph

-0.64

uden

-0.64

elight

-0.62

 widest

-0.62

POSITIVE LOGITS

 member

1.11

 participant

0.92

 subscriber

0.91

 descendant

0.86

 resident

0.84

 registered

0.83

verse

0.83

 supporter

0.82

 follower

0.81

 believer

0.81

Activations Density 0.059%

a specific string related to verifying humanity, often found in a captcha-like context asking users to confirm they are not robots

instances of the article "a" in various contexts

No Comments

No Known Activations