INDEX

Explanations

notes or warnings within text

oai_token-act-pair · gpt-3.5-turbo

occurrences of notes or annotations within the text

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

undai

-0.79

izons

-0.78

ivable

-0.78

wre

-0.73

ailability

-0.73

uce

-0.71

 elim

-0.70

eatures

-0.69

 wreck

-0.68

 tremend

-0.68

POSITIVE LOGITS

TBD

0.87

 Unable

0.81

 Provided

0.76

 Exactly

0.76

 Previous

0.75

 Originally

0.74

 When

0.73

 Correct

0.73

 Beware

0.73

*)

0.72

Activations Density 0.071%

notes or warnings within text

occurrences of notes or annotations within the text

No Comments

No Known Activations