INDEX

Explanations

instances where individuals or groups are urging or strongly recommending something

oai_token-act-pair · gpt-3.5-turbo

instances of the word "urging."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

issance

-0.64

JJ

-0.63

fy

-0.63

idental

-0.61

Category

-0.61

fit

-0.60

ossal

-0.59

ys

-0.58

ibles

-0.58

phabet

-0.58

POSITIVE LOGITS

 urging

3.38

 advising

1.86

 pleading

1.75

 recommending

1.72

 reminding

1.66

 advocating

1.59

 denouncing

1.56

 asking

1.51

 thanking

1.48

 praising

1.48

Activations Density 0.017%

instances where individuals or groups are urging or strongly recommending something

instances of the word "urging."

No Comments

No Known Activations