INDEX

Explanations

symbols '&' and the word 'and'

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

owe

-0.71

stood

-0.71

soDeliveryDate

-0.71

lake

-0.69

ograp

-0.68

hiba

-0.67

natureconservancy

-0.67

seless

-0.67

sylv

-0.66

lain

-0.65

POSITIVE LOGITS

romeda

0.66

 Torres

0.65

 Herz

0.63

Ãī

0.63

ãĥ¼ãĥĨ

0.61

APTER

0.61

ãģª

0.61

 furthermore

0.61

âĶģ

0.60

brates

0.60

Activations Density 0.151%

symbols '&' and the word 'and'

No Comments

No Known Activations