INDEX

Explanations

words related to physical or emotional suffering or hardship

oai_token-act-pair · gpt-3.5-turbo

references to the concept of "grueling" experiences or conditions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

 Doctrine

-0.67

 FISA

-0.66

itect

-0.63

 Izan

-0.62

 Mirror

-0.62

 harbor

-0.61

 Surveillance

-0.60

 IMAGES

-0.59

 Partners

-0.59

 Perception

-0.58

POSITIVE LOGITS

ppo

1.00

gru

0.97

eling

0.96

ppy

0.93

ache

0.89

¸

0.89

esome

0.88

grim

0.85

ppe

0.83

arse

0.83

Activations Density 0.016%

words related to physical or emotional suffering or hardship

references to the concept of "grueling" experiences or conditions

No Comments

No Known Activations