INDEX

Explanations

parenthetical statements that contain a variety of information or context

oai_token-act-pair · gpt-3.5-turbo

opening parentheses

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 spir

-0.88

 undet

-0.87

 accrued

-0.84

 prepar

-0.83

 stocking

-0.83

 clin

-0.83

 intensive

-0.83

 accumulated

-0.82

 tackling

-0.82

 overwhelmed

-0.80

POSITIVE LOGITS

See

1.60

Note

1.53

Spoiler

1.51

Though

1.49

Actually

1.47

Answer

1.47

Side

1.45

Remember

1.45

Unless

1.44

Ironically

1.44

Activations Density 0.051%

parenthetical statements that contain a variety of information or context

opening parentheses

No Comments

No Known Activations