INDEX

Explanations

words related to providing information or updating others

oai_token-act-pair · gpt-3.5-turbo

instances of the word "inform."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

âĻ¥

-0.68

FUN

-0.65

Kings

-0.65

BR

-0.65

Dub

-0.62

engeance

-0.62

 vain

-0.62

DAY

-0.61

mad

-0.61

BUR

-0.61

POSITIVE LOGITS

tale

1.02

ally

0.97

 inform

0.93

ingly

0.91

ations

0.85

acion

0.81

ially

0.80

sheets

0.79

ificantly

0.79

 Subscribe

0.76

Activations Density 0.008%

words related to providing information or updating others

instances of the word "inform."

No Comments

No Known Activations