INDEX

Explanations

text related to the concept of "bits"

oai_token-act-pair · gpt-3.5-turbo

the repetition of the term "bits" 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 @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

ammad

-0.69

inate

-0.69

NOR

-0.68

inated

-0.65

SOS

-0.61

 ceiling

-0.61

 Physicians

-0.60

rouse

-0.60

zanne

-0.59

INE

-0.58

POSITIVE LOGITS

 Bits

1.21

 bits

1.21

bits

1.21

terness

0.98

coins

0.91

wagon

0.91

wana

0.89

bread

0.85

layer

0.84

buck

0.83

Activations Density 0.010%

text related to the concept of "bits"

the repetition of the term "bits" in various contexts

No Comments

No Known Activations