INDEX

Explanations

people's names, locations, and organizations

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

manship

-0.71

pload

-0.70

kins

-0.69

heads

-0.67

head

-0.65

ptive

-0.64

Ø©

-0.64

STON

-0.64

kers

-0.63

gb

-0.62

POSITIVE LOGITS

ral

1.25

ropy

1.20

rip

1.13

een

1.12

ribution

1.12

orian

1.12

rained

1.12

rum

1.11

rations

1.10

uary

1.10

Activations Density 2.498%

people's names, locations, and organizations

No Comments

No Known Activations