INDEX

Explanations

mentions or references to former individuals or things

oai_token-act-pair · gpt-3.5-turbo

the prefix "ex-" indicating former or previous associations or statuses

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

 Sabha

-0.91

 Flavoring

-0.82

 GOODMAN

-0.79

ãĤ¤ãĥĪ

-0.78

jriwal

-0.73

GoldMagikarp

-0.71

WAYS

-0.69

HCR

-0.68

OOD

-0.68

veyard

-0.68

POSITIVE LOGITS

uber

1.10

clamation

1.09

ogenous

1.06

orbit

0.97

terior

0.95

oplan

0.95

clud

0.94

uding

0.91

uded

0.89

tern

0.87

Activations Density 0.009%

mentions or references to former individuals or things

the prefix "ex-" indicating former or previous associations or statuses

No Comments

No Known Activations