INDEX

Explanations

phrases related to elimination or removal

oai_token-act-pair · gpt-3.5-turbo

phrases related to processes of elimination

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

fo

-0.67

ificent

-0.64

Alert

-0.64

chester

-0.64

 gulf

-0.62

Ub

-0.61

 Ples

-0.61

Attach

-0.60

cit

-0.59

ourning

-0.59

POSITIVE LOGITS

rase

3.16

 elimination

1.99

 Elim

1.12

ersen

1.07

verson

1.06

ioxide

1.05

issors

1.04

inav

1.01

plet

1.01

rika

0.94

Activations Density 0.056%

phrases related to elimination or removal

phrases related to processes of elimination

No Comments

No Known Activations