INDEX

Explanations

names or words containing "wer"

oai_token-act-pair · gpt-3.5-turbo

occurrences of the sequence "wer" 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 @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

ortium

-0.75

ividual

-0.67

 reinforcement

-0.62

EStream

-0.60

ãģĨ

-0.59

 Coyotes

-0.57

CPC

-0.57

 narrated

-0.56

UAL

-0.56

 Labrador

-0.56

POSITIVE LOGITS

ijn

1.02

wer

0.99

riter

0.99

rer

0.91

pees

0.86

tyard

0.85

pee

0.84

ijk

0.83

nesday

0.83

ij

0.81

Activations Density 0.006%

names or words containing "wer"

occurrences of the sequence "wer" in various contexts

No Comments

No Known Activations