INDEX

Explanations

morphic words or terms

oai_token-act-pair · gpt-3.5-turbo

descriptors related to size, particularly "medium" and its modifiers

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

atche

-0.75

gments

-0.71

dfx

-0.67

AMES

-0.67

apons

-0.64

ADRA

-0.63

printed

-0.63

ATHER

-0.62

phia

-0.62

 Merch

-0.62

POSITIVE LOGITS

 sized

1.35

weights

0.92

weight

0.92

sized

0.83

ship

0.81

 thickness

0.81

 term

0.79

 length

0.78

 size

0.77

 range

0.75

Activations Density 0.040%

morphic words or terms

descriptors related to size, particularly "medium" and its modifiers

No Comments

No Known Activations