INDEX

Explanations

terms related to capabilities or the potential of something

oai_token-act-pair · gpt-3.5-turbo

references to various skills or functions of systems or technologies

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

orno

-0.76

birth

-0.74

eree

-0.70

idan

-0.67

boys

-0.66

cott

-0.65

paying

-0.64

esta

-0.64

hound

-0.63

apest

-0.63

POSITIVE LOGITS

 capabilities

1.00

 capability

0.89

 conferred

0.85

BILITIES

0.85

abilities

0.82

hered

0.79

Ability

0.77

 abilities

0.77

 afforded

0.76

mentation

0.73

Activations Density 0.055%

terms related to capabilities or the potential of something

references to various skills or functions of systems or technologies

No Comments

No Known Activations