INDEX

Explanations

proper nouns or names related to people or places

oai_token-act-pair · gpt-3.5-turbo

names and terms associated with groups or organization structures

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 offenses

-0.76

 inmate

-0.71

 charact

-0.71

 unden

-0.67

Charges

-0.65

 offense

-0.63

ruciating

-0.62

 welf

-0.62

itability

-0.61

DAQ

-0.61

POSITIVE LOGITS

ordon

1.40

naire

0.96

roup

0.91

arding

0.84

ordial

0.79

ieu

0.77

-+-+-+-+

0.77

 largeDownload

0.77

éĹĺ

0.77

enhagen

0.75

Activations Density 0.019%

proper nouns or names related to people or places

names and terms associated with groups or organization structures

No Comments

No Known Activations