INDEX

Explanations

words in a different language, potentially related to an error or a different character encoding

oai_token-act-pair · gpt-3.5-turbo

sequences of non-standard or encoded characters, possibly relating to different languages

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 bearer

-0.64

 Thomson

-0.52

 authenticity

-0.50

 underdog

-0.48

 nonpartisan

-0.47

acebook

-0.47

 authentic

-0.46

 Broadcasting

-0.46

 confidential

-0.46

 negotiator

-0.45

POSITIVE LOGITS

"></

0.69

Ø©

0.66

)).

0.64

¶

0.61

²

0.60

scl

0.59

ensis

0.59

}.

0.59

nova

0.59

Į

0.57

Activations Density 0.390%

words in a different language, potentially related to an error or a different character encoding

sequences of non-standard or encoded characters, possibly relating to different languages

No Comments

No Known Activations