INDEX

Explanations

instances where the word "ill" appears in the text

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "this" and variations of "ill."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

lished

-0.73

 Shapiro

-0.70

imov

-0.64

 bestowed

-0.63

ISON

-0.60

 Petr

-0.59

 repro

-0.58

 Cancel

-0.58

 Pixie

-0.57

 Chester

-0.57

POSITIVE LOGITS

ohyd

0.85

ogether

0.77

opot

0.76

axis

0.76

asty

0.76

agra

0.75

ewater

0.74

 Nadu

0.74

meal

0.71

cture

0.70

Activations Density 0.048%

instances where the word "ill" appears in the text

occurrences of the word "this" and variations of "ill."

No Comments

No Known Activations