INDEX

Explanations

references to leaks or leakage

oai_token-act-pair · gpt-3.5-turbo

references to leaks or the concept of leaking

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

iterranean

-0.79

oran

-0.79

aple

-0.72

adv

-0.72

uliffe

-0.71

merce

-0.71

ommel

-0.71

aldi

-0.69

nesota

-0.68

arching

-0.68

POSITIVE LOGITS

 leak

1.29

 leaks

1.17

Leaks

1.12

 leaking

1.08

 leakage

0.87

 leaked

0.84

 valves

0.80

inventoryQuantity

0.78

ileaks

0.76

 valve

0.74

Activations Density 0.006%

references to leaks or leakage

references to leaks or the concept of leaking

No Comments

No Known Activations