INDEX

Explanations

phrases containing the word "fault"

oai_token-act-pair · gpt-3.5-turbo

references to faults or errors in various contexts

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ago

-0.77

xy

-0.74

¥µ

-0.74

het

-0.74

sov

-0.73

erva

-0.72

atten

-0.71

rooms

-0.70

iolet

-0.69

rons

-0.69

POSITIVE LOGITS

 fault

1.49

 faults

1.17

 Fault

1.05

 forgiven

0.99

 blame

0.88

 excuse

0.76

 blames

0.73

 tolerant

0.72

 forgiveness

0.70

 tolerance

0.69

Activations Density 0.006%

phrases containing the word "fault"

references to faults or errors in various contexts

No Comments

No Known Activations