INDEX

Explanations

keywords related to rectifying, fixing, or correcting something

oai_token-act-pair · gpt-3.5-turbo

terms related to rectification processes and corrections

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

UGH

-0.77

ARK

-0.70

WAYS

-0.68

PLIED

-0.68

ESH

-0.65

ÙĲ

-0.64

EEK

-0.63

lihood

-0.63

Ø§ÙĦ

-0.62

 Resistance

-0.61

POSITIVE LOGITS

angles

1.12

 rect

1.09

rect

1.03

ified

1.02

angular

0.99

uitous

0.98

rices

0.97

ifying

0.95

ifiable

0.95

itude

0.92

Activations Density 0.014%

keywords related to rectifying, fixing, or correcting something

terms related to rectification processes and corrections

No Comments

No Known Activations