INDEX

Explanations

specific phrases indicating progression in a text

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate conditional or hypothetical scenarios

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

PF

-0.67

MPG

-0.63

ãĥĨ

-0.63

ãĥ¼ãĥĨ

-0.60

ASED

-0.59

ectomy

-0.58

âĢ¢âĢ¢âĢ¢âĢ¢

-0.58

Interstitial

-0.57

 Primary

-0.57

 Marble

-0.57

POSITIVE LOGITS

ratulations

0.77

aucus

0.75

oiler

0.72

ften

0.69

 congr

0.69

endor

0.68

deck

0.68

 parting

0.67

yah

0.67

hew

0.66

Activations Density 0.145%

specific phrases indicating progression in a text

phrases that indicate conditional or hypothetical scenarios

No Comments

No Known Activations