INDEX

Explanations

instances of numerical years

oai_token-act-pair · gpt-3.5-turbo

references to the year 2010

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

Magikarp

-0.78

ournals

-0.78

ques

-0.77

amins

-0.75

esses

-0.73

ess

-0.72

phy

-0.71

wat

-0.68

ournal

-0.68

kell

-0.64

POSITIVE LOGITS

å¹

1.00

 census

0.83

acious

0.78

 onwards

0.75

 Census

0.74

 Affordable

0.72

ullah

0.72

 Remastered

0.69

ford

0.69

 Tigers

0.68

Activations Density 0.016%

instances of numerical years

references to the year 2010

No Comments

No Known Activations