INDEX

Explanations

names of places or entities

oai_token-act-pair · gpt-3.5-turbo

references to examples and comparisons

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

 imperialist

-0.85

ascus

-0.82

 Whilst

-0.78

ocaust

-0.77

respective

-0.77

 Sovere

-0.76

secution

-0.75

 Conclusion

-0.74

ocide

-0.74

Therefore

-0.69

POSITIVE LOGITS

 brainstorm

0.80

 Craigslist

0.77

 perk

0.74

 Yelp

0.74

 Pinterest

0.74

 broch

0.73

 favorites

0.72

 classmate

0.72

 personalized

0.72

 webcam

0.71

Activations Density 1.778%

names of places or entities

references to examples and comparisons

No Comments

No Known Activations