INDEX

Explanations

names or events related to a specific historical context or setting

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_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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

é¾įå

-0.66

 Covenant

-0.66

ãģ®éŃĶ

-0.65

hetical

-0.64

ILA

-0.64

 Unch

-0.63

ographical

-0.62

 Defenders

-0.62

EMS

-0.61

enger

-0.61

POSITIVE LOGITS

iosity

1.25

tain

1.14

few

1.14

rencies

1.11

rier

1.08

iously

1.08

mud

1.07

rying

1.04

ves

1.02

ricular

1.02

Activations Density 0.019%

names or events related to a specific historical context or setting

No Comments

No Known Activations