INDEX

Explanations

acronyms and codes with numbers in them

oai_token-act-pair · gpt-3.5-turbo

references to specific bills or legislative measures

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

hoff

-0.78

poke

-0.72

etsk

-0.68

 Feast

-0.65

nings

-0.64

 caution

-0.62

ensional

-0.62

lift

-0.62

isite

-0.61

edIn

-0.60

POSITIVE LOGITS

ulhu

0.86

emporary

0.80

ottage

0.74

section

0.74

redible

0.72

Section

0.71

sections

0.70

WINDOWS

0.69

0.68

Users

0.68

Activations Density 0.042%

acronyms and codes with numbers in them

references to specific bills or legislative measures

No Comments

No Known Activations