INDEX

Explanations

words related to locations or places being "home" to something or someone

oai_token-act-pair · gpt-3.5-turbo 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

sha

-0.72

params

-0.60

uku

-0.59

ussian

-0.59

ional

-0.57

 haircut

-0.57

asm

-0.57

 stride

-0.56

err

-0.56

 undert

-0.56

POSITIVE LOGITS

coming

0.98

brew

0.91

liest

0.90

grown

0.85

lier

0.83

chool

0.79

oided

0.79

opathy

0.79

quartered

0.78

pun

0.77

Activations Density 0.034%

words related to locations or places being "home" to something or someone

No Comments

No Known Activations