INDEX

Explanations

references to locations, specifically related to a place called "Garden."

oai_token-act-pair · gpt-3.5-turbo

references to "Garden" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

 derail

-0.77

senal

-0.77

DonaldTrump

-0.71

ifles

-0.71

aundering

-0.71

etimes

-0.70

merce

-0.69

worldly

-0.69

ickr

-0.65

esty

-0.65

POSITIVE LOGITS

 Garden

1.08

hire

0.96

 Grove

0.87

arden

0.86

 Bowl

0.86

ertodd

0.86

keeper

0.83

hill

0.82

 Awards

0.79

esville

0.78

Activations Density 0.007%

references to locations, specifically related to a place called "Garden."

references to "Garden" in various contexts

No Comments

No Known Activations