INDEX

Explanations

references to locations or geographical entities, particularly with the abbreviation "NE."

oai_token-act-pair · gpt-3.5-turbo

references to geographical locations, particularly in Nebraska

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

acies

-0.73

doms

-0.71

loo

-0.69

framework

-0.65

acters

-0.64

indal

-0.63

avorite

-0.62

aults

-0.62

rate

-0.62

gerald

-0.61

POSITIVE LOGITS

VE

0.97

ITH

0.89

IGH

0.86

VEN

0.84

erd

0.83

FU

0.79

ISS

0.79

meric

0.79

OM

0.77

JM

0.77

Activations Density 0.017%

references to locations or geographical entities, particularly with the abbreviation "NE."

references to geographical locations, particularly in Nebraska

No Comments

No Known Activations