INDEX

Explanations

mentions of specific types of structures, specifically buildings

oai_token-act-pair · gpt-3.5-turbo

terms related to buildings and construction

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

nir

-0.79

Ã±o

-0.73

sha

-0.71

sen

-0.69

owder

-0.68

nia

-0.67

romeda

-0.65

tle

-0.64

pas

-0.64

 Nadu

-0.64

POSITIVE LOGITS

 buildings

1.18

raper

1.08

 Buildings

1.05

chool

0.96

 skysc

0.95

 blocks

0.91

 erected

0.89

 complexes

0.89

 constructed

0.84

 towers

0.83

Activations Density 0.019%

mentions of specific types of structures, specifically buildings

terms related to buildings and construction

No Comments

No Known Activations