INDEX

Explanations

phrases related to things never seen or done before

oai_token-act-pair · gpt-3.5-turbo

references to experiences or events that have never been encountered previously

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

area

-0.64

ãĥ¼ãĥ

-0.62

atana

-0.57

Ĳ

-0.57

ãĤ¤

-0.55

 mart

-0.55

©

-0.54

opoulos

-0.54

inn

-0.54

omo

-0.54

POSITIVE LOGITS

heed

0.73

hene

0.71

orthy

0.68

htaking

0.68

NetMessage

0.66

bnb

0.65

nas

0.64

isner

0.62

 fading

0.61

LOD

0.60

Activations Density 0.031%

phrases related to things never seen or done before

references to experiences or events that have never been encountered previously

No Comments

No Known Activations