INDEX

Explanations

references to counter-related concepts or activities

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "counter" 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 @ 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

 Forge

-0.79

ABE

-0.78

 Known

-0.75

doms

-0.73

 Pelicans

-0.72

OOD

-0.72

 Roche

-0.72

 Tornado

-0.70

 Cursed

-0.68

 Pistons

-0.67

POSITIVE LOGITS

measures

1.23

intuitive

1.16

attack

1.13

intelligence

1.13

balance

1.10

counter

1.09

offensive

1.03

ror

1.00

fact

1.00

balanced

0.98

Activations Density 0.016%

references to counter-related concepts or activities

occurrences of the word "counter" in various contexts

No Comments

No Known Activations