INDEX

Explanations

the word "nearest"

oai_token-act-pair · gpt-3.5-turbo

the concept of proximity or nearness

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

anity

-0.81

inse

-0.79

raft

-0.74

zinski

-0.72

skill

-0.72

eri

-0.72

era

-0.70

detail

-0.70

ocker

-0.69

ulet

-0.68

POSITIVE LOGITS

 nearest

1.17

 closest

0.87

 neighb

0.84

 neighbour

0.84

 nearer

0.83

terness

0.81

Ĭ±

0.80

oreal

0.79

ĸļ

0.78

 fart

0.76

Activations Density 0.005%

the word "nearest"

the concept of proximity or nearness

No Comments

No Known Activations