INDEX

Explanations

phrases related to expressing statements or claims

oai_token-act-pair · gpt-3.5-turbo

phrases and words that indicate reported speech or claims made by individuals or organizations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

ratulations

-0.72

ntil

-0.72

earances

-0.65

abi

-0.63

estern

-0.62

itiveness

-0.61

Reply

-0.61

ption

-0.60

tan

-0.59

izons

-0.58

POSITIVE LOGITS

 constitutes

0.99

 represents

0.98

 violates

0.91

 proves

0.89

 amounted

0.89

 comprises

0.83

 undermines

0.82

 signifies

0.80

 resembles

0.79

 embodies

0.79

Activations Density 0.085%

phrases related to expressing statements or claims

phrases and words that indicate reported speech or claims made by individuals or organizations

No Comments

No Known Activations