INDEX

Explanations

phrases related to affirmations or confirmations

oai_token-act-pair · gpt-3.5-turbo

instances of direct speech or quotations

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

 afterlife

-0.79

 purse

-0.79

 totem

-0.79

 pudding

-0.77

 mosqu

-0.77

 contested

-0.74

 quickest

-0.73

 concess

-0.71

 fermented

-0.71

 extinguished

-0.70

POSITIVE LOGITS

 Then

0.96

 Asked

0.93

 Later

0.89

 "...

0.87

 Meanwhile

0.86

 Afterwards

0.86

 That

0.84

 Exactly

0.84

 However

0.84

 "âĢ¦

0.84

Activations Density 0.166%

phrases related to affirmations or confirmations

instances of direct speech or quotations

No Comments

No Known Activations