INDEX

Explanations

mentions of specific locations, governmental positions, and job titles

oai_token-act-pair · gpt-3.5-turbo

geographical and organizational entities, particularly related to military and political contexts

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

IUM

-0.68

ternity

-0.65

teness

-0.65

paces

-0.64

 mathemat

-0.63

SHIP

-0.61

 Faust

-0.58

ships

-0.57

hops

-0.57

 Shard

-0.57

POSITIVE LOGITS

oreAnd

0.67

 director

0.66

 analyst

0.62

 attorney

0.62

vet

0.61

 practition

0.61

 pione

0.61

riter

0.60

arella

0.59

ationally

0.59

Activations Density 0.974%

mentions of specific locations, governmental positions, and job titles

geographical and organizational entities, particularly related to military and political contexts

No Comments

No Known Activations