INDEX

Explanations

statements where confirmation is provided

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "confirmed" in various contexts

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

 ILCS

-0.75

assic

-0.73

asus

-0.71

brance

-0.68

EStreamFrame

-0.67

Pg

-0.66

enium

-0.65

oiler

-0.64

phabet

-0.63

ptives

-0.63

POSITIVE LOGITS

 receipt

1.02

 that

0.84

 suspicions

0.82

 rumours

0.79

 rumors

0.78

via

0.75

 sightings

0.73

to

0.70

 sighting

0.70

 anecd

0.70

Activations Density 0.043%

statements where confirmation is provided

occurrences of the word "confirmed" in various contexts

No Comments

No Known Activations