INDEX

Explanations

names of individuals or organizations with slight variations

oai_token-act-pair · gpt-3.5-turbo

sequences involving proper nouns or names

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

ulhu

-0.75

 Misc

-0.71

lasses

-0.70

SOS

-0.66

 nomine

-0.66

 fuck

-0.63

 FIGHT

-0.63

LW

-0.63

[+]

-0.61

 replay

-0.61

POSITIVE LOGITS

Hart

0.95

Birth

0.92

Marie

0.91

Pont

0.91

Hol

0.89

Mer

0.88

Pal

0.86

Cent

0.86

Web

0.86

Charl

0.85

Activations Density 0.079%

names of individuals or organizations with slight variations

sequences involving proper nouns or names

No Comments

No Known Activations