INDEX

Explanations

statistics and percentages in text

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_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

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

sort

-0.69

abl

-0.65

spir

-0.63

alph

-0.62

ful

-0.61

pedia

-0.61

andal

-0.59

hement

-0.59

 Cosponsors

-0.58

help

-0.58

POSITIVE LOGITS

GDP

1.01

 population

0.94

 respondents

0.92

 variance

0.91

 households

0.88

 total

0.84

 surveyed

0.83

 earners

0.83

 Americans

0.81

 overall

0.81

Activations Density 1.349%

statistics and percentages in text

No Comments

No Known Activations