INDEX

Explanations

words related to emotional responses or feedback

oai_token-act-pair · gpt-3.5-turbo

phrases related to responses or reactions to events or situations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

hold

-0.71

ramer

-0.70

ffe

-0.68

locked

-0.68

fusc

-0.65

enture

-0.64

orah

-0.63

inav

-0.63

 mortg

-0.62

 Sinai

-0.62

POSITIVE LOGITS

 reaction

1.24

 reactions

1.22

ivation

1.09

ivated

1.05

 Reaction

1.05

aries

1.02

ivating

0.95

naires

0.88

naire

0.85

 responses

0.85

Activations Density 0.018%

words related to emotional responses or feedback

phrases related to responses or reactions to events or situations

No Comments

No Known Activations