INDEX

Explanations

phrases indicating lack of completion or action

oai_token-act-pair · gpt-3.5-turbo

the phrase "haven't" and its variants 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

dress

-0.69

otype

-0.67

upp

-0.67

urat

-0.67

rophe

-0.59

yles

-0.58

mens

-0.57

 guid

-0.57

 Ballistic

-0.56

angering

-0.56

POSITIVE LOGITS

't

1.09

ited

0.89

ÃŃ

0.86

 Been

0.84

tyard

0.83

geon

0.82

anu

0.81

ģĸ

0.80

dayName

0.79

eteen

0.79

Activations Density 0.015%

phrases indicating lack of completion or action

the phrase "haven't" and its variants in various contexts

No Comments

No Known Activations