INDEX

Explanations

phrases related to conflict or legal matters

oai_token-act-pair · gpt-3.5-turbo

references to the concept of "trauma" or "traumatic experiences"

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

 Mellon

-0.65

hyde

-0.63

ensable

-0.62

ãģ®éŃĶ

-0.61

ENSE

-0.60

iod

-0.59

 cooler

-0.58

manship

-0.58

ERN

-0.58

heit

-0.58

POSITIVE LOGITS

pez

0.99

tra

0.97

ppings

0.96

ph

0.92

vert

0.92

ordinary

0.90

tro

0.88

alog

0.84

inen

0.84

rend

0.83

Activations Density 0.005%

phrases related to conflict or legal matters

references to the concept of "trauma" or "traumatic experiences"

No Comments

No Known Activations