INDEX

Explanations

comparisons between different quantities or attributes

oai_token-act-pair · gpt-3.5-turbo

instances of comparison

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

authorized

-0.73

STAR

-0.61

 Polo

-0.59

 Permanent

-0.59

 Publication

-0.58

 Semin

-0.58

gren

-0.57

OTA

-0.57

Bay

-0.57

wheel

-0.55

POSITIVE LOGITS

 favorably

0.96

lihood

0.82

 thereto

0.73

 Compare

0.73

icut

0.71

Compare

0.68

 compare

0.66

 disadvant

0.66

illac

0.64

ileaks

0.64

Activations Density 0.018%

comparisons between different quantities or attributes

instances of comparison

No Comments

No Known Activations