INDEX

Explanations

specific references to the term "Homer" with varying levels of activation

oai_token-act-pair · gpt-3.5-turbo

instances of the name "Homer" and its variations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

pots

-0.72

birth

-0.65

PER

-0.65

²¾

-0.65

Ĥª

-0.63

 Assassins

-0.62

comings

-0.61

orld

-0.60

Ws

-0.60

CLOSE

-0.60

POSITIVE LOGITS

astics

0.84

cial

0.82

ic

0.81

arch

0.80

anz

0.80

archy

0.78

ufact

0.76

ocl

0.76

ophone

0.76

andom

0.76

Activations Density 0.030%

specific references to the term "Homer" with varying levels of activation

instances of the name "Homer" and its variations

No Comments

No Known Activations