INDEX

Explanations

text snippets containing the word "se" followed by a number

oai_token-act-pair · gpt-3.5-turbo

words related to restraint or confinement

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 Herb

-0.63

 Banner

-0.61

Uni

-0.60

SHIP

-0.60

UD

-0.60

 Pence

-0.59

ï¸ı

-0.59

 Nights

-0.59

 Union

-0.59

ington

-0.59

POSITIVE LOGITS

ptic

1.31

gue

1.26

perate

1.26

pta

1.23

xt

1.23

rene

1.20

rend

1.18

pt

1.18

xton

1.13

cluded

1.12

Activations Density 0.015%

text snippets containing the word "se" followed by a number

words related to restraint or confinement

No Comments

No Known Activations