INDEX

Explanations

phrases related to surprise or revelation

oai_token-act-pair · gpt-3.5-turbo

statements related to surprising or unexpected information and their phrasing

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

ascript

-0.68

ourses

-0.66

sequently

-0.65

ilial

-0.64

rans

-0.62

[+

-0.61

[_

-0.61

azel

-0.61

Applications

-0.58

emort

-0.57

POSITIVE LOGITS

 kidding

0.89

 nerds

0.77

 damned

0.75

 darn

0.75

 understatement

0.75

 Cheap

0.73

 steroids

0.72

 hats

0.71

 Dirty

0.68

 classy

0.68

Activations Density 1.991%

phrases related to surprise or revelation

statements related to surprising or unexpected information and their phrasing

No Comments

No Known Activations