INDEX

Explanations

potential or possibility expressed through the word "could"

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

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

-0.72

Maker

-0.67

 Duty

-0.65

bound

-0.64

paced

-0.60

 Updated

-0.59

 honors

-0.59

holder

-0.59

posts

-0.58

making

-0.58

POSITIVE LOGITS

 feas

1.29

 conce

1.18

 hypot

1.08

 theoretically

1.06

 possibly

0.98

ivably

0.95

 easily

0.94

 argue

0.91

 alternatively

0.91

've

0.88

Activations Density 0.083%

potential or possibility expressed through the word "could"

No Comments

No Known Activations