INDEX

Explanations

numbers with a plus sign preceding them

oai_token-act-pair · gpt-3.5-turbo

numerical values associated with statistical or performance metrics

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

 Sutherland

-0.74

©¶æ

-0.74

onite

-0.72

vey

-0.71

anium

-0.70

ãĤ©

-0.70

 Winc

-0.70

 Grimes

-0.69

ning

-0.68

 Stub

-0.68

POSITIVE LOGITS

/-

1.25

 IMAGES

0.89

-+

0.78

/+

0.77

plus

0.75

ordinary

0.73

cffffcc

0.73

ishly

0.72

#$

0.72

scl

0.71

Activations Density 0.008%

numbers with a plus sign preceding them

numerical values associated with statistical or performance metrics

No Comments

No Known Activations