INDEX

Explanations

phrases indicating potential outcomes or implications

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate meaning and consequences

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

ĸļ

-0.75

Omn

-0.67

Jur

-0.67

 Wend

-0.66

 Britann

-0.64

raint

-0.63

Jal

-0.61

 Sabb

-0.61

Defense

-0.58

 Herz

-0.58

POSITIVE LOGITS

depending

0.78

ESE

0.72

gettable

0.72

 depending

0.71

Î³

0.70

pole

0.70

 tricky

0.70

 safely

0.69

 easily

0.68

 GOODMAN

0.68

Activations Density 0.293%

phrases indicating potential outcomes or implications

phrases that indicate meaning and consequences

No Comments

No Known Activations