INDEX

Explanations

instances where something is capable of being done or has the ability to do something

oai_token-act-pair · gpt-3.5-turbo

the repetition of the word "could" 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 @ 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

core

-0.70

pires

-0.70

furt

-0.66

 Duty

-0.64

 Appears

-0.64

balls

-0.63

 Updated

-0.62

holder

-0.62

 Meeting

-0.61

houses

-0.60

POSITIVE LOGITS

 feas

1.22

 afford

1.10

 conce

1.05

 muster

0.97

 possibly

0.95

've

0.93

 easily

0.90

nt

0.90

 foresee

0.89

 argue

0.88

Activations Density 0.094%

instances where something is capable of being done or has the ability to do something

the repetition of the word "could" in various contexts

No Comments

No Known Activations