INDEX

Explanations

lines of code or technical commands written in a specific format

oai_token-act-pair · gpt-3.5-turbo

equality and assignment operations in coding or mathematical contexts

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

 braces

-0.71

zynski

-0.70

ergy

-0.69

 resume

-0.63

eding

-0.63

 establishment

-0.63

 breed

-0.62

algia

-0.61

anni

-0.60

wcs

-0.60

POSITIVE LOGITS

ãĥīãĥ©ãĤ´ãĥ³

1.04

PsyNetMessage

0.96

CVE

0.92

0.88

saf

0.85

ãĥ´ãĤ¡

0.84

true

0.84

mc

0.83

imm

0.82

mint

0.82

Activations Density 0.013%

lines of code or technical commands written in a specific format

equality and assignment operations in coding or mathematical contexts

No Comments

No Known Activations