INDEX

Explanations

references to a specific name or term "ante"

oai_token-act-pair · gpt-3.5-turbo

mentions of the term "ante" in various contexts

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

nesota

-0.94

stones

-0.83

anyl

-0.80

dfx

-0.78

abies

-0.77

stone

-0.75

matically

-0.74

soDeliveryDate

-0.73

loss

-0.72

lik

-0.71

POSITIVE LOGITS

zzi

0.95

Ã«

0.84

rer

0.81

ctr

0.77

Ã±o

0.77

ante

0.77

zza

0.77

chnology

0.75

uve

0.75

ques

0.73

Activations Density 0.017%

references to a specific name or term "ante"

mentions of the term "ante" in various contexts

No Comments

No Known Activations