INDEX

Explanations

adjectives and nouns related to importance or impact

oai_token-act-pair · gpt-3.5-turbo

key adjectives or descriptors that indicate significant qualities or conditions

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

ĸļ

-1.08

ICLE

-0.87

 Debor

-0.73

heid

-0.67

lain

-0.65

dq

-0.64

 TAMADRA

-0.64

%]

-0.64

olutely

-0.63

 Uriel

-0.62

POSITIVE LOGITS

igans

0.72

worldly

0.70

 flair

0.70

uitous

0.66

else

0.66

place

0.66

away

0.65

legged

0.65

 tricks

0.63

 whatsoever

0.63

Activations Density 0.197%

adjectives and nouns related to importance or impact

key adjectives or descriptors that indicate significant qualities or conditions

No Comments

No Known Activations