INDEX

Explanations

terms related to specific locations, activities, and names of individuals

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

UNCH

-0.67

ARP

-0.62

IPP

-0.62

XT

-0.61

ELF

-0.61

FORE

-0.60

ASED

-0.60

ICT

-0.59

CE

-0.58

astern

-0.56

POSITIVE LOGITS

2.75

sburg

1.71

sie

1.54

sis

1.51

sb

1.44

ski

1.40

sburgh

1.38

sf

1.37

sin

1.34

sch

1.34

Activations Density 3.903%

terms related to specific locations, activities, and names of individuals

No Comments

No Known Activations