INDEX

Explanations

political terms or references

oai_token-act-pair · gpt-3.5-turbo

terms related to politics and political discussions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Shining

-0.75

 Warrant

-0.75

 warrant

-0.71

 recall

-0.71

 Carbuncle

-0.70

Ved

-0.68

 Norn

-0.68

 warrants

-0.67

urat

-0.64

 Covenant

-0.60

POSITIVE LOGITS

icians

1.70

ician

1.58

ifact

1.52

ically

1.36

eness

1.34

ico

1.33

icans

1.18

ican

1.12

icker

1.10

ique

1.09

Activations Density 0.035%

political terms or references

terms related to politics and political discussions

No Comments

No Known Activations