INDEX

Explanations

words related to specific parts or sections within a larger context

oai_token-act-pair · gpt-3.5-turbo

references to "bits."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

WAY

-0.70

ammad

-0.69

inated

-0.69

NOR

-0.65

inate

-0.64

 Physicians

-0.63

INE

-0.61

Bal

-0.61

SOS

-0.61

WAYS

-0.60

POSITIVE LOGITS

terness

1.05

 Bits

1.00

wagon

1.00

 bits

0.96

bits

0.94

itives

0.89

umen

0.88

meal

0.84

bread

0.81

buck

0.81

Activations Density 0.008%

words related to specific parts or sections within a larger context

references to "bits."

No Comments

No Known Activations