INDEX

Explanations

action-related terms or verbs related to various activities

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ifies

-0.78

iosis

-0.77

ngth

-0.77

idon

-0.76

overe

-0.74

uate

-0.72

uria

-0.72

fi

-0.71

uther

-0.71

adian

-0.71

POSITIVE LOGITS

 mechanism

0.86

rod

0.78

 point

0.76

 ability

0.74

 circle

0.73

 spree

0.72

 agent

0.71

 Stones

0.71

 force

0.71

 grounds

0.70

Activations Density 0.190%

action-related terms or verbs related to various activities

No Comments

No Known Activations