INDEX

Explanations

words related to finality or conclusions

oai_token-act-pair · gpt-3.5-turbo

phrases that emphasize finality or conclusions

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

Wee

-0.68

Normally

-0.65

 Recently

-0.60

maid

-0.59

kh

-0.59

 Frog

-0.59

RN

-0.59

 Flying

-0.59

 formerly

-0.58

 regularly

-0.58

POSITIVE LOGITS

 unsuccessful

0.79

 culmin

0.79

 deciding

0.79

 decide

0.77

 succeed

0.77

apor

0.75

 succeeded

0.73

 decided

0.73

 succumb

0.72

 boils

0.70

Activations Density 0.028%

words related to finality or conclusions

phrases that emphasize finality or conclusions

No Comments

No Known Activations