INDEX

Explanations

terms related to additional or extra resources or support

oai_token-act-pair · gpt-3.5-turbo

terms related to supplemental or auxiliary information

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

rant

-0.78

shake

-0.75

âĸĪâĸĪâĸĪâĸĪ

-0.72

reth

-0.70

washer

-0.70

jar

-0.70

âĸĪâĸĪ

-0.70

 Wilde

-0.70

Cola

-0.69

riz

-0.68

POSITIVE LOGITS

 supplemental

1.16

 Supplemental

1.10

 supplementary

0.95

 auxiliary

0.92

endum

0.89

hesis

0.78

 subp

0.75

 thereto

0.73

ufficient

0.72

 aids

0.72

Activations Density 0.019%

terms related to additional or extra resources or support

terms related to supplemental or auxiliary information

No Comments

No Known Activations