INDEX

Explanations

references to interactions or relationships with neighbors

oai_token-act-pair · gpt-3.5-turbo

mentions of neighbors or neighboring entities

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

inen

-0.89

ebook

-0.82

uchin

-0.76

oa

-0.75

anche

-0.74

obe

-0.74

BN

-0.73

othal

-0.73

nosis

-0.72

obin

-0.71

POSITIVE LOGITS

 neighbors

1.43

 neighbours

1.18

 neighbor

1.14

 Neigh

1.09

 neighbour

0.93

 neighb

0.87

Neigh

0.85

 Neighborhood

0.85

bors

0.84

 roomm

0.83

Activations Density 0.006%

references to interactions or relationships with neighbors

mentions of neighbors or neighboring entities

No Comments

No Known Activations