INDEX

Explanations

words related to caution, warning, or advisory messages

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

Negative Logits

erity

-0.92

ngth

-0.91

iewicz

-0.87

jc

-0.82

eq

-0.81

heast

-0.81

atem

-0.80

leased

-0.79

ilty

-0.79

duino

-0.79

POSITIVE LOGITS

 charm

0.90

 hatch

0.86

 pill

0.83

 hawk

0.83

 drill

0.81

 circle

0.81

 tale

0.79

 blaster

0.78

 basket

0.77

 Throne

0.77

Activations Density 0.232%

words related to caution, warning, or advisory messages

No Known Activations