INDEX

Explanations

keywords related to names or references to specific names, especially "Dar"

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

sburgh

-0.75

ħĭ

-0.75

 actionGroup

-0.73

éĹĺ

-0.71

ATK

-0.64

 nomine

-0.63

VALUE

-0.63

IDs

-0.63

 seism

-0.62

 Skydragon

-0.61

POSITIVE LOGITS

lington

1.19

geon

1.09

rah

1.03

ques

1.00

izon

0.97

acle

0.96

ien

0.95

rington

0.93

rian

0.93

fur

0.92

Activations Density 0.016%

keywords related to names or references to specific names, especially "Dar"

No Comments

No Known Activations