INDEX

Explanations

mentions of notable individuals or achievements, such as climbing Mount Everest or winning a Pulitzer, especially related to achievements or medical conditions

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_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

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

-0.79

frame

-0.75

emate

-0.70

lopp

-0.70

erker

-0.66

ÑĮ

-0.66

hani

-0.66

abases

-0.65

ouver

-0.65

fram

-0.64

POSITIVE LOGITS

 Falls

0.67

 flats

0.62

 Haram

0.59

halla

0.58

 Labs

0.57

 Alley

0.56

 Herz

0.56

 annex

0.54

 markings

0.53

 Parish

0.53

Activations Density 3.908%

mentions of notable individuals or achievements, such as climbing Mount Everest or winning a Pulitzer, especially related to achievements or medical conditions

No Comments

No Known Activations