INDEX

Explanations

negations and denials in speech

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res_post_32k-oai

Configuration

jbloom/GPT2-Small-OAI-v5-32k-resid-post-SAEs/v5_32k_layer_6.pt

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

32,768

Data Type

torch.float32

Hook Name

blocks.6.hook_resid_post

Hook Layer

Architecture

standard

Context Size

Dataset

Skylion007/openwebtext

Activation Function

topk

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ilts

-0.71

çļ

-0.71

ò

-0.66

 pione

-0.66

ñ

-0.65

 Greenberg

-0.64

ĪĴ

-0.64

 guiActiveUnfocused

-0.64

©¶æ

-0.64

 Estate

-0.63

POSITIVE LOGITS

psy

0.63

etus

0.62

ctic

0.62

ct

0.61

ling

0.61

perate

0.59

eni

0.58

 prospect

0.58

 conscious

0.57

 starved

0.57

Activations Density 0.138%

negations and denials in speech

No Comments

No Known Activations