INDEX

Explanations

phrases indicating joking or exaggeration

oai_token-act-pair · gpt-3.5-turbo

expressions indicating humor or sarcasm

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

CTR

-0.64

roots

-0.59

 Bridges

-0.58

iliated

-0.58

 manned

-0.56

reens

-0.56

icipated

-0.56

 intersect

-0.56

 scrim

-0.56

 competed

-0.55

POSITIVE LOGITS

^^^^

0.86

_.

0.85

 kidding

0.82

 haha

0.79

;)

0.79

:-)

0.76

 myself

0.75

:)

0.74

idge

0.73

 here

0.72

Activations Density 0.403%

phrases indicating joking or exaggeration

expressions indicating humor or sarcasm

No Comments

No Known Activations