INDEX

Explanations

numerical values in a specific format

oai_token-act-pair · gpt-3.5-turbo

occurrences of the number "1" 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 @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

awaru

-0.73

 hygiene

-0.68

 propri

-0.62

alus

-0.62

azel

-0.59

 mascul

-0.57

ategory

-0.57

orative

-0.57

hift

-0.57

aimon

-0.56

POSITIVE LOGITS

st

1.63

1.15

1.10

Password

1.05

1.03

1.02

0.94

0.85

ST

0.84

 sqor

0.83

Activations Density 0.098%

numerical values in a specific format

occurrences of the number "1" in various contexts

No Comments

No Known Activations