INDEX

Explanations

phrases indicating the compilation or organization of information

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

 externalToEVAOnly

-0.67

engers

-0.63

ipal

-0.63

 Canaveral

-0.63

[+

-0.60

ofi

-0.59

 Pradesh

-0.59

å£«

-0.59

obyl

-0.59

Charges

-0.58

POSITIVE LOGITS

 rundown

1.18

 infographic

1.17

 below

0.96

 handy

0.94

 roundup

0.94

 curated

0.93

Below

0.93

 HuffPost

0.89

 Below

0.89

 summarize

0.87

Activations Density 0.520%

phrases indicating the compilation or organization of information

No Comments

No Known Activations