INDEX

Explanations

references to recent events or 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_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

claimer

-0.77

amina

-0.68

olor

-0.65

 presently

-0.64

ike

-0.63

 Directorate

-0.62

requires

-0.62

pose

-0.61

 Binary

-0.60

pit

-0.59

POSITIVE LOGITS

 spate

0.94

å¹

0.89

 afternoon

0.82

ãĥ¼ãĥ³

0.81

 developments

0.81

 evenings

0.77

 evening

0.77

 controversies

0.76

 morning

0.75

 additions

0.74

Activations Density 0.373%

references to recent events or activities

No Comments

No Known Activations