INDEX

Explanations

mentions of objects or remains being discovered in various locations

oai_token-act-pair · gpt-3.5-turbo

references to discoveries of deceased individuals or remains

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.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

Controller

-0.72

VIDIA

-0.70

spons

-0.70

yout

-0.69

thood

-0.68

 assumes

-0.67

annon

-0.67

 announcer

-0.67

!/

-0.67

ocity

-0.65

POSITIVE LOGITS

 archaeological

1.04

 Treasure

0.89

 roadside

0.86

 excavation

0.86

 wreckage

0.85

 trove

0.84

 landfill

0.84

 attic

0.84

 vicinity

0.82

 woods

0.82

Activations Density 0.275%

mentions of objects or remains being discovered in various locations

references to discoveries of deceased individuals or remains

No Comments

No Known Activations