INDEX

Explanations

locations or objects related to trains

oai_token-act-pair · gpt-3.5-turbo

references to train services and transportation

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

osi

-0.71

 Sphere

-0.66

 Reach

-0.66

 Tablet

-0.66

erenn

-0.65

monds

-0.64

 Prairie

-0.63

cised

-0.61

ÏĤ

-0.61

pring

-0.60

POSITIVE LOGITS

wreck

1.29

 conductor

1.06

 derail

1.05

loads

1.03

 wreck

1.02

ees

1.01

 ride

1.00

 passenger

0.99

riages

0.98

roads

0.96

Activations Density 0.031%

locations or objects related to trains

references to train services and transportation

No Comments

No Known Activations