INDEX

Explanations

terms related to historical or political forms of government power, specifically focused on rules and rulers

oai_token-act-pair · gpt-3.5-turbo

references to forms of governance or authority

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

ãĥ¤

-0.73

velength

-0.70

 reviewer

-0.69

actory

-0.67

 Sense

-0.67

 sqor

-0.64

reci

-0.64

 lymph

-0.63

 molecular

-0.62

 linem

-0.62

POSITIVE LOGITS

doms

0.84

 presided

0.84

pin

0.81

hip

0.75

orship

0.73

esses

0.72

dom

0.72

edom

0.71

thood

0.71

pins

0.70

Activations Density 0.052%

terms related to historical or political forms of government power, specifically focused on rules and rulers

references to forms of governance or authority

No Comments

No Known Activations