INDEX

Explanations

names and terms related to specific entities or individuals

oai_token-act-pair · gpt-3.5-turbo

names of people or characters

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

 multim

-0.73

 nationally

-0.58

 lottery

-0.58

 comparable

-0.56

 academ

-0.55

enegger

-0.54

 smack

-0.54

SPONSORED

-0.53

 psychiat

-0.53

upt

-0.53

POSITIVE LOGITS

idae

0.89

ania

0.88

ak

0.88

osaurus

0.87

acus

0.87

ius

0.85

ka

0.85

Å«

0.84

ai

0.84

obia

0.83

Activations Density 0.232%

names and terms related to specific entities or individuals

names of people or characters

No Comments

No Known Activations