INDEX

Explanations

references to specific locations or events

oai_token-act-pair · gpt-3.5-turbo

identifiable subjects and objects in a range of contexts

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

ãĥĥãĥī

-0.86

"âĢ¦

-0.82

CLASS

-0.74

 outright

-0.72

Ļ½

-0.70

Again

-0.69

Wow

-0.69

"?

-0.69

CONCLUS

-0.68

"—

-0.68

POSITIVE LOGITS

window

0.97

 Shutterstock

0.85

 illustration

0.82

 skyline

0.81

 silhou

0.81

 sunrise

0.79

 logo

0.74

 makeshift

0.74

 wreckage

0.73

FILE

0.72

Activations Density 0.328%

references to specific locations or events

identifiable subjects and objects in a range of contexts

No Comments

No Known Activations