INDEX

Explanations

a variety of prepositions and conjunctions that imply relationships or connections

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

Negative Logits

Recomm

-0.64

 Inquis

-0.63

Lago

-0.62

è¦ļéĨĴ

-0.61

 prayers

-0.61

BIL

-0.60

 showc

-0.60

IAS

-0.59

 rapt

-0.59

Limited

-0.58

POSITIVE LOGITS

hemy

1.10

etooth

0.81

haven

0.78

steen

0.77

itzer

0.75

ĵĺ

0.74

zinski

0.74

velt

0.72

keye

0.72

geoning

0.72

Activations Density 0.037%

a variety of prepositions and conjunctions that imply relationships or connections

No Known Activations