INDEX

Explanations

technical instructions related to wiring and connections in hardware or software systems

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

å§«

-0.79

otropic

-0.75

Saharan

-0.61

Ļ½

-0.61

romy

-0.60

foreign

-0.60

£ı

-0.60

 FANTASY

-0.59

ometric

-0.59

idious

-0.59

POSITIVE LOGITS

 itself

1.07

 portion

0.75

 alone

0.74

 themselves

0.73

quo

0.71

 closest

0.69

 Quart

0.65

eers

0.65

 afore

0.64

izes

0.64

Activations Density 0.431%

technical instructions related to wiring and connections in hardware or software systems

No Comments

No Known Activations