INDEX

Explanations

words related to locations or settings, particularly those involving events or incidents

oai_token-act-pair · gpt-3.5-turbo

repetitions of the word "scene"

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

utm

-0.75

rand

-0.75

ledge

-0.71

hips

-0.69

limited

-0.64

gain

-0.63

gran

-0.63

iam

-0.60

Increases

-0.60

 elected

-0.58

POSITIVE LOGITS

 scene

4.03

 Scene

2.87

scene

2.72

 scenes

2.66

 Scenes

2.04

Scene

2.03

scenes

1.50

 scenery

1.16

 sequence

1.07

 scen

1.05

Activations Density 0.027%

words related to locations or settings, particularly those involving events or incidents

repetitions of the word "scene"

No Comments

No Known Activations