INDEX

Explanations

numbers or alphanumeric codes embedded within text

oai_token-act-pair · gpt-3.5-turbo

numerical identifiers associated with products or content

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 millenn

-0.84

 nomine

-0.73

 hearty

-0.72

reau

-0.68

lings

-0.66

 suprem

-0.64

 historians

-0.64

 scrut

-0.63

 mayoral

-0.63

neys

-0.63

POSITIVE LOGITS

 partName

1.19

1.07

1.04

1.03

Activations Density 0.046%

numbers or alphanumeric codes embedded within text

numerical identifiers associated with products or content

No Comments

No Known Activations