INDEX

Explanations

words related to assigning or evaluating values or scores to entities

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_fs12288-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_12288

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

12,288

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

DOC

-0.74

aptic

-0.71

loo

-0.68

sil

-0.67

bane

-0.66

bite

-0.66

bush

-0.65

ENG

-0.63

 combustion

-0.63

FORE

-0.62

POSITIVE LOGITS

 allegiance

0.84

 blame

0.82

 assignments

0.81

 assigns

0.79

 assignment

0.79

ments

0.77

 assigning

0.76

 duties

0.75

 tasks

0.74

 permissions

0.74

Activations Density 0.082%

words related to assigning or evaluating values or scores to entities

No Comments

No Known Activations