INDEX

Explanations

powerful or impactful concepts and ideas

oai_token-act-pair · gpt-3.5-turbo

terms and phrases that denote significance or impact

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

çīĪ

-0.77

İĭ

-0.75

ORTS

-0.72

Xi

-0.70

ADRA

-0.69

lehem

-0.69

ONDON

-0.69

OULD

-0.69

ħĭ

-0.66

aucuses

-0.66

POSITIVE LOGITS

 commodity

0.84

 sacrament

0.81

 predictor

0.79

 causation

0.77

 everywhere

0.77

 thing

0.76

 scarce

0.76

 mistress

0.72

 motiv

0.71

 Sacrament

0.71

Activations Density 0.507%

powerful or impactful concepts and ideas

terms and phrases that denote significance or impact

No Comments

No Known Activations