INDEX

Explanations

terms related to organizations or proper nouns abbreviated as "NC"

oai_token-act-pair · gpt-3.5-turbo

references to organizations or institutions abbreviated as "NC" followed by numerical designations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

thing

-0.91

chest

-0.90

loo

-0.89

tons

-0.86

gerald

-0.76

shake

-0.73

nings

-0.72

Picture

-0.70

SourceFile

-0.69

_____

-0.68

POSITIVE LOGITS

urses

1.05

ODE

0.90

STAR

0.86

ouri

0.83

ADE

0.79

ategor

0.79

ERT

0.77

urrent

0.73

urse

0.73

ossus

0.73

Activations Density 0.013%

terms related to organizations or proper nouns abbreviated as "NC"

references to organizations or institutions abbreviated as "NC" followed by numerical designations

No Comments

No Known Activations