INDEX

Explanations

instances of "what" followed by a comparison or contrast in the text

oai_token-act-pair · gpt-3.5-turbo

instances of the word "what" 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 @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

enburg

-0.69

ster

-0.65

eer

-0.62

jee

-0.62

robe

-0.61

ature

-0.60

adden

-0.59

Sao

-0.59

 Gamb

-0.58

isher

-0.58

POSITIVE LOGITS

 happened

1.15

 happens

1.15

soever

1.12

 transpired

1.02

 happ

0.96

 constitutes

0.94

 kinds

0.90

 constituted

0.87

 sorts

0.85

 else

0.83

Activations Density 0.105%

instances of "what" followed by a comparison or contrast in the text

instances of the word "what" in various contexts

No Comments

No Known Activations