INDEX

Explanations

instances of the word "relatively" and related phrases of moderation or comparison

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

 Polo

-0.79

olor

-0.71

 Cities

-0.71

ourses

-0.71

rea

-0.70

andel

-0.69

ved

-0.69

VL

-0.68

halla

-0.66

 Orchestra

-0.66

POSITIVE LOGITS

 tame

1.04

 unpop

1.01

 benign

0.96

 insignificant

0.96

 unaffected

0.95

 harmless

0.95

 uncont

0.93

 inexpensive

0.92

 innocuous

0.92

 insensitive

0.91

Activations Density 0.026%

instances of the word "relatively" and related phrases of moderation or comparison

No Comments

No Known Activations