INDEX

Explanations

positive words or phrases, possibly related to recommendations or praise

oai_token-act-pair · gpt-3.5-turbo

the expression of strong opinions or recommendations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

olicy

-0.67

IPM

-0.65

 Everton

-0.61

office

-0.61

evidence

-0.60

 âĢİ

-0.59

 Oswald

-0.58

bars

-0.58

nature

-0.58

 Abortion

-0.58

POSITIVE LOGITS

cknowled

1.81

cknow

1.58

verages

1.14

HAHAHAHA

1.13

HAHA

1.11

HHHH

1.05

chie

1.04

lyss

0.97

UTH

0.95

roma

0.95

Activations Density 0.106%

positive words or phrases, possibly related to recommendations or praise

the expression of strong opinions or recommendations

No Comments

No Known Activations