INDEX

Explanations

mentions of objects or people bending or being bent

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

gado

-0.66

 Carbuncle

-0.65

Reporting

-0.61

 Lauder

-0.60

pta

-0.60

DAQ

-0.60

calling

-0.60

 Scotia

-0.59

 Countdown

-0.59

 Sunny

-0.59

POSITIVE LOGITS

¿½

0.96

 knees

0.88

strings

0.84

ĸļ

0.84

 bends

0.83

bent

0.80

 bend

0.79

ĺħ

0.78

 bent

0.77

onite

0.75

Activations Density 0.029%

mentions of objects or people bending or being bent

No Comments

No Known Activations