INDEX

Explanations

references to the letter "H" followed by a number, possibly related to a specific entity or product

oai_token-act-pair · gpt-3.5-turbo

mentions of the letter 'H' followed by a number

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

ãĥ¼ãĥĨãĤ£

-0.84

enhagen

-0.76

DragonMagazine

-0.74

 totality

-0.71

EStream

-0.71

 Clicker

-0.71

 Beware

-0.70

 Conquest

-0.68

ãĥ¯

-0.66

CPR

-0.66

POSITIVE LOGITS

ospital

1.10

awk

1.10

IGH

1.08

OST

1.06

acking

1.04

umble

1.04

aven

1.03

OLD

1.02

olly

1.02

annah

1.01

Activations Density 0.044%

references to the letter "H" followed by a number, possibly related to a specific entity or product

mentions of the letter 'H' followed by a number

No Comments

No Known Activations