INDEX

Explanations

words related to correction or improvement

oai_token-act-pair · gpt-3.5-turbo

terms related to correction and adjustments

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

ARK

-0.70

Ø§ÙĦ

-0.68

 Resistance

-0.66

è¦ļéĨĴ

-0.64

UGH

-0.62

 Prosper

-0.60

ÙĲ

-0.59

 resisting

-0.58

PLIED

-0.58

âĸ¬âĸ¬

-0.58

POSITIVE LOGITS

angles

1.07

angular

1.07

uitous

1.00

itude

0.98

ifiable

0.96

rect

0.91

 rect

0.90

racted

0.87

anus

0.86

ions

0.86

Activations Density 0.038%

words related to correction or improvement

terms related to correction and adjustments

No Comments

No Known Activations