INDEX

Explanations

medical terms and possibly technical or programming-related terms

oai_token-act-pair · gpt-3.5-turbo

references to specific items or entities within a narrative context

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

nikov

-0.62

zai

-0.58

jit

-0.56

 Kavanaugh

-0.54

 Tsarnaev

-0.53

 Murdoch

-0.52

Assad

-0.52

heid

-0.52

strom

-0.51

 Lauder

-0.50

POSITIVE LOGITS

its

0.63

 uniqueness

0.62

 Himself

0.61

 fulfill

0.58

 Everyday

0.57

iverse

0.57

 anytime

0.55

 itself

0.54

ITS

0.54

heres

0.54

Activations Density 1.263%

medical terms and possibly technical or programming-related terms

references to specific items or entities within a narrative context

No Comments

No Known Activations