INDEX

Explanations

phrases indicating a summary or conclusion

oai_token-act-pair · gpt-3.5-turbo

phrases indicating collective experiences or shared situations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

ror

-0.72

 anymore

-0.71

 whisper

-0.61

lev

-0.61

gey

-0.60

opp

-0.59

Ã§a

-0.58

 panic

-0.58

":""},{"

-0.58

drm

-0.57

POSITIVE LOGITS

 totals

0.93

Ĥª

0.83

 totaling

0.82

ogether

0.81

 total

0.78

 comprise

0.77

 amassed

0.75

 suffice

0.74

 tidy

0.73

 roughly

0.72

Activations Density 0.571%

phrases indicating a summary or conclusion

phrases indicating collective experiences or shared situations

No Comments

No Known Activations