INDEX

Explanations

references to sources or authority figures in statements

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs49152-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

49,152

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

 farm

-0.64

llular

-0.64

blast

-0.64

pez

-0.63

ql

-0.61

igers

-0.61

 heights

-0.61

Osc

-0.60

 realise

-0.58

 roller

-0.58

POSITIVE LOGITS

ly

0.90

edly

0.79

eous

0.74

views

0.73

Ĥİ

0.73

ĪĴ

0.71

tained

0.71

sequently

0.70

dylib

0.69

translation

0.69

Activations Density 0.026%

references to sources or authority figures in statements

No Comments

No Known Activations