INDEX

Explanations

phrases related to specific actions or tasks that need to be completed

oai_token-act-pair · gpt-3.5-turbo

references to specific settings or contexts within narratives

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

ities

-0.76

 channels

-0.68

 piles

-0.66

iencies

-0.61

lees

-0.61

bats

-0.60

annels

-0.60

 establishments

-0.59

riages

-0.59

inas

-0.59

POSITIVE LOGITS

 titled

0.67

lished

0.62

 consisting

0.59

slot

0.58

İĭ

0.57

 resembling

0.57

pload

0.57

ressor

0.56

ogram

0.56

racuse

0.56

Activations Density 0.839%

phrases related to specific actions or tasks that need to be completed

references to specific settings or contexts within narratives

No Comments

No Known Activations