INDEX

Explanations

verbs that indicate or suggest information

oai_token-act-pair · gpt-3.5-turbo

words that express indications or signals regarding an event or condition

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

@#&

-0.90

fare

-0.88

ctors

-0.83

zan

-0.73

iling

-0.68

venge

-0.68

pping

-0.68

ney

-0.65

ÄŁ

-0.65

tal

-0.64

POSITIVE LOGITS

 indications

0.95

 signs

0.91

ered

0.85

manship

0.81

 hints

0.80

 Signs

0.79

 indicates

0.79

 indication

0.78

 intervals

0.78

 indicated

0.75

Activations Density 0.038%

verbs that indicate or suggest information

words that express indications or signals regarding an event or condition

No Comments

No Known Activations