INDEX

Explanations

names of products or items commonly found in a specific setting or context

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_fs3072-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

3,072

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

 Scots

-0.72

ACTED

-0.71

 foremost

-0.69

 Columb

-0.63

izations

-0.61

TAIN

-0.61

 Scandinavian

-0.60

 conspicuous

-0.60

ndra

-0.59

ritional

-0.57

POSITIVE LOGITS

ombie

1.37

arro

1.25

hou

1.25

alez

1.13

ombies

1.07

hao

1.05

arella

1.03

ymes

0.99

quez

0.99

akura

0.92

Activations Density 0.937%

names of products or items commonly found in a specific setting or context

No Comments

No Known Activations