INDEX

Explanations

words related to activities or actions on physical objects, potentially with a focus on technical or mechanical actions

oai_token-act-pair · gpt-3.5-turbo

references to actions and experiences related to personal involvement or events

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

Â¯Â¯Â¯Â¯

-0.74

 Prosecut

-0.60

 themselves

-0.60

void

-0.60

Engineers

-0.59

 Juven

-0.57

 sidelines

-0.57

 liabilities

-0.57

 blight

-0.56

 delim

-0.56

POSITIVE LOGITS

 myself

1.20

poke

0.76

EStream

0.72

miah

0.71

 fortunate

0.68

undai

0.67

 sugg

0.66

teness

0.66

saw

0.64

my

0.64

Activations Density 1.356%

words related to activities or actions on physical objects, potentially with a focus on technical or mechanical actions

references to actions and experiences related to personal involvement or events

No Comments

No Known Activations