INDEX

Explanations

symbols and representations related to data or information

oai_token-act-pair · gpt-3.5-turbo

terms related to representation and symbolism in graphics and visual communication

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.5.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

alm

-0.77

kees

-0.72

nown

-0.69

BILL

-0.69

sbm

-0.65

wage

-0.63

ppa

-0.63

uss

-0.62

aughter

-0.62

kick

-0.61

POSITIVE LOGITS

 everything

0.73

Interstitial

0.67

 them

0.65

ANCE

0.65

 individuality

0.65

 warmth

0.64

"+

0.63

ments

0.63

 continuity

0.63

us

0.61

Activations Density 0.319%

symbols and representations related to data or information

terms related to representation and symbolism in graphics and visual communication

No Comments

No Known Activations