INDEX

Explanations

information and evidence related to data and accounts

oai_token-act-pair · gpt-3.5-turbo

references to evidence, data, and accounts related to reports and analysis

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

amaz

-0.71

ilet

-0.66

 importance

-0.66

 jeopardy

-0.66

rection

-0.65

osion

-0.64

illard

-0.64

ãĥ¤

-0.63

 Dragonbound

-0.63

verning

-0.63

POSITIVE LOGITS

 collected

1.35

 supplied

1.25

 gathered

1.24

 glean

1.23

 obtained

1.22

 extracted

1.15

 compiled

1.13

 drawn

1.12

 derived

1.06

 taken

1.04

Activations Density 0.385%

information and evidence related to data and accounts

references to evidence, data, and accounts related to reports and analysis

No Comments

No Known Activations