INDEX

Explanations

expressions related to strong emotional reactions or outbursts, especially negative ones

oai_token-act-pair · gpt-3.5-turbo

language reflecting strong negative emotions and aggressive expressions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

ipeg

-0.78

bley

-0.77

itions

-0.76

igm

-0.74

itional

-0.74

ernels

-0.74

flix

-0.73

fman

-0.72

yles

-0.71

lymp

-0.70

POSITIVE LOGITS

tir

1.10

 insults

0.93

 vitri

0.93

 comments

0.92

 outburst

0.89

 slurs

0.84

 commenters

0.80

 spew

0.80

 tant

0.79

 yells

0.79

Activations Density 0.038%

expressions related to strong emotional reactions or outbursts, especially negative ones

language reflecting strong negative emotions and aggressive expressions

No Comments

No Known Activations