INDEX

Explanations

technical terms related to software, programming, and commands

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

ricular

-0.75

à¨

-0.71

å§«

-0.67

POSE

-0.67

 Instruments

-0.64

xious

-0.63

CLASSIFIED

-0.61

ificant

-0.61

 Brow

-0.60

ordinate

-0.59

POSITIVE LOGITS

ilogy

0.74

 wagon

0.72

undrum

0.69

naires

0.67

 fiasco

0.67

owder

0.66

ifies

0.66

naire

0.65

ea

0.65

lance

0.65

Activations Density 2.350%

technical terms related to software, programming, and commands

No Comments

No Known Activations