INDEX

Explanations

words ending in 'in' which may refer to names or specific locations

oai_token-act-pair · gpt-3.5-turbo

instances of the word "in" across various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

 Seym

-0.84

 warr

-0.80

 destro

-0.76

 redund

-0.75

 behav

-0.71

 payday

-0.71

ģĸ

-0.71

hovah

-0.70

 entrants

-0.69

 trave

-0.68

POSITIVE LOGITS

strument

1.43

aug

1.25

jury

1.22

iti

1.20

ners

1.16

flation

1.13

jured

1.13

clusive

1.11

clus

1.11

ned

1.09

Activations Density 0.052%

words ending in 'in' which may refer to names or specific locations

instances of the word "in" across various contexts

No Comments

No Known Activations