INDEX

Explanations

phrases containing the specific initials "HM" followed by a number

oai_token-act-pair · gpt-3.5-turbo

references to specific military or naval entities

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

heimer

-0.81

feed

-0.76

eve

-0.68

margin

-0.65

 Buccaneers

-0.65

 Machina

-0.65

byss

-0.63

eus

-0.62

anza

-0.60

lect

-0.60

POSITIVE LOGITS

HM

0.91

DOM

0.85

emonic

0.84

ICAN

0.83

abyte

0.78

ITED

0.75

puter

0.74

awei

0.73

cipl

0.73

IME

0.72

Activations Density 0.015%

phrases containing the specific initials "HM" followed by a number

references to specific military or naval entities

No Comments

No Known Activations