INDEX

Explanations

numerical values, especially those denoted with a specific formatting including "Â±" and "cm/s"

oai_token-act-pair · gpt-3.5-turbo

statistical values or measurements presented with uncertainty

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

neys

-0.86

arians

-0.73

unciation

-0.72

arium

-0.72

aco

-0.71

zon

-0.68

 Citiz

-0.68

aries

-0.68

arius

-0.65

iger

-0.64

POSITIVE LOGITS

=~=~

0.91

/-

0.88

/+

0.86

phies

0.86

cffffcc

0.84

¥µ

0.81

Â±

0.76

+/-

0.75

hap

0.71

ÃĽ

0.71

Activations Density 0.014%

numerical values, especially those denoted with a specific formatting including "Â±" and "cm/s"

statistical values or measurements presented with uncertainty

No Comments

No Known Activations