INDEX

Explanations

references to citations or sources within a text

oai_token-act-pair · gpt-3.5-turbo

references to structured formats or lists, often indicated by brackets or ellipses

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

 favour

-0.66

 herself

-0.64

 governors

-0.64

oria

-0.62

 community

-0.61

 synthes

-0.61

adra

-0.61

 foreseeable

-0.61

uria

-0.60

azes

-0.60

POSITIVE LOGITS

ONSORED

1.00

However

0.84

Its

0.82

Also

0.81

They

0.81

Sources

0.80

Therefore

0.78

Unless

0.76

hide

0.75

Compare

0.74

Activations Density 0.129%

references to citations or sources within a text

references to structured formats or lists, often indicated by brackets or ellipses

No Comments

No Known Activations