INDEX

Explanations

words related to pointing out or emphasizing important aspects or details

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

uthor

-0.77

ESE

-0.76

OOL

-0.73

DAQ

-0.72

natureconservancy

-0.70

ctic

-0.65

 Seym

-0.64

ccording

-0.63

ÄŁ

-0.63

cling

-0.63

POSITIVE LOGITS

lessly

0.91

 Reyes

0.90

 deduction

0.88

blank

0.88

point

0.84

 pointing

0.83

iasis

0.83

points

0.82

 blank

0.82

 deductions

0.79

Activations Density 1.951%

words related to pointing out or emphasizing important aspects or details

No Comments

No Known Activations