INDEX

Explanations

phrases related to physical paths like trails

oai_token-act-pair · gpt-3.5-turbo

references to trails and outdoor pathways

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

entric

-0.76

rators

-0.71

ricular

-0.67

rontal

-0.64

oln

-0.64

ities

-0.62

ascular

-0.61

berra

-0.61

ivot

-0.61

 subsidized

-0.61

POSITIVE LOGITS

 trail

1.23

 trails

1.02

 Blazers

0.96

ĸļ

0.83

 Trail

0.81

walker

0.80

 Runner

0.80

runner

0.79

gate

0.79

runners

0.78

Activations Density 0.011%

phrases related to physical paths like trails

references to trails and outdoor pathways

No Comments

No Known Activations