INDEX

Explanations

keywords related to reactions and responses

oai_token-act-pair · gpt-3.5-turbo

instances of the word "reaction."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

inav

-0.79

©¶æ

-0.74

locked

-0.74

 mortg

-0.70

 sliced

-0.67

estern

-0.67

anto

-0.64

pil

-0.64

 carp

-0.63

aper

-0.63

POSITIVE LOGITS

 reaction

1.49

 reactions

1.48

 Reaction

1.37

ivation

1.00

ivated

0.89

aries

0.86

 responses

0.85

 response

0.84

 react

0.81

 retaliate

0.80

Activations Density 0.009%

keywords related to reactions and responses

instances of the word "reaction."

No Comments

No Known Activations