INDEX

Explanations

references to specific sci-fi franchises and related terminology

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_fs3072-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

3,072

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

odcast

-0.81

ipop

-0.78

utics

-0.78

ĵĺ

-0.75

Downloadha

-0.75

ãĥ¼ãĥ³

-0.73

Ń·

-0.72

ecause

-0.70

ĸļ

-0.69

LIA

-0.69

POSITIVE LOGITS

burst

1.11

light

1.08

bucks

1.04

vation

1.01

ring

0.98

buck

0.95

fish

0.95

stru

0.93

lit

0.90

fighter

0.88

Activations Density 0.667%

references to specific sci-fi franchises and related terminology

No Comments

No Known Activations