INDEX

Explanations

names of locations, particularly with the word "Hills" in them

oai_token-act-pair · gpt-3.5-turbo

references to geographical locations, specifically hills and boroughs

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

istically

-0.94

sonian

-0.81

bilt

-0.80

thumbnails

-0.78

istics

-0.76

binary

-0.73

isms

-0.73

*/(

-0.72

istic

-0.72

netic

-0.71

POSITIVE LOGITS

borough

1.22

dale

1.04

creen

0.98

peria

0.96

boro

0.92

onse

0.84

earch

0.75

ourced

0.74

pring

0.74

hee

0.73

Activations Density 0.083%

names of locations, particularly with the word "Hills" in them

references to geographical locations, specifically hills and boroughs

No Comments

No Known Activations