INDEX

Explanations

occurrences of specific prefixes or suffixes related to various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

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

thodox

-0.79

ramid

-0.77

drm

-0.76

idays

-0.67

plement

-0.64

izons

-0.62

earances

-0.62

azaki

-0.62

ancial

-0.62

yip

-0.61

POSITIVE LOGITS

agues

0.68

 Rasm

0.67

å°Ĩ

0.67

 Phillip

0.66

osterone

0.64

ãĥ¢

0.62

 Disease

0.62

 Gillespie

0.61

Rae

0.61

adelphia

0.60

Activations Density 0.056%

occurrences of specific prefixes or suffixes related to various contexts

No Comments

No Known Activations