INDEX

Explanations

specific patterns or phrases associated with alternative or contrasting actions

oai_token-act-pair · gpt-3.5-turbo

phrases indicating assumptions or alternatives

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

stad

-0.73

 compr

-0.70

stead

-0.69

berra

-0.66

SG

-0.64

DL

-0.64

IQ

-0.63

culosis

-0.62

hard

-0.61

DA

-0.60

POSITIVE LOGITS

Downloadha

0.73

SPONSORED

0.72

mere

0.71

amiya

0.69

 ours

0.68

 Instead

0.68

 relying

0.68

 recourse

0.67

isites

0.67

Ľ

0.67

Activations Density 0.084%

specific patterns or phrases associated with alternative or contrasting actions

phrases indicating assumptions or alternatives

No Comments

No Known Activations