INDEX

Explanations

phrases related to representing or symbolizing something

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate representation or indicate the act of representing something or someone

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

 liking

-0.65

 hurry

-0.65

 curfew

-0.62

 twitch

-0.60

 CCTV

-0.57

 quick

-0.57

 busy

-0.56

 manual

-0.55

bern

-0.55

 checking

-0.54

POSITIVE LOGITS

 represents

3.45

 represent

2.30

 signifies

2.14

 represented

2.13

 embodies

2.04

 corresponds

1.89

 representing

1.87

 denotes

1.84

 constitutes

1.79

resents

1.78

Activations Density 0.025%

phrases related to representing or symbolizing something

phrases that indicate representation or indicate the act of representing something or someone

No Comments

No Known Activations