INDEX

Explanations

instances where references are made to being transparent about mistakes, including grammar-related errors and personal errors

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Configuration

jbloom/Gemma-2b-IT-Residual-Stream-SAEs/gemma_2b_it_blocks.12.hook_resid_post_16384

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

HuggingFaceFW/fineweb

Features

16,384

Data Type

float32

Hook Name

blocks.12.hook_resid_post

Hook Layer

Architecture

standard

Context Size

1,024

Dataset

Skylion007/openwebtext

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Majest

-0.74

vnt

-0.70

 addirittura

-0.66

 impra

-0.66

Juf

-0.65

fta

-0.65

 Augu

-0.64

 thut

-0.64

 délib

-0.63

 Bibl

-0.62

POSITIVE LOGITS

<bos>

0.80

 inevitable

0.65

 inevitably

0.65

 sometime

0.60

 occasional

0.59

 occasionally

0.56

 eventually

0.56

SneakyThrows

0.54

CodedInputStream

0.53

 somewhere

0.50

Activations Density 0.367%

instances where references are made to being transparent about mistakes, including grammar-related errors and personal errors

No Comments

No Known Activations