INDEX

Explanations

words related to geographical locations, specifically countries and cities

oai_token-act-pair · gpt-3.5-turbo

mentions of geographic locations, particularly countries and cities

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

RAM

-0.77

manship

-0.76

ota

-0.73

worthiness

-0.69

unal

-0.68

lez

-0.67

alog

-0.67

oom

-0.66

ork

-0.64

RNA

-0.63

POSITIVE LOGITS

ãĤ¶

0.80

bledon

0.80

uates

0.75

enthal

0.69

 submar

0.69

odies

0.68

yip

0.67

cellence

0.67

iddler

0.67

 Aires

0.66

Activations Density 0.048%

words related to geographical locations, specifically countries and cities

mentions of geographic locations, particularly countries and cities

No Comments

No Known Activations