INDEX

Explanations

words related to text formatting, such as 'AND' likely related to section headings or emphasis

oai_token-act-pair · gpt-3.5-turbo

the conjunction "and" 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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

nces

-0.71

fitting

-0.69

kes

-0.65

kers

-0.65

ãĤ·ãĥ£

-0.65

ches

-0.65

ched

-0.64

housing

-0.64

natureconservancy

-0.61

ctions

-0.61

POSITIVE LOGITS

AND

1.19

romeda

0.98

ARD

0.93

OTO

0.88

OVER

0.87

 THEN

0.86

AY

0.84

IDE

0.84

AL

0.83

ILL

0.82

Activations Density 0.017%

words related to text formatting, such as 'AND' likely related to section headings or emphasis

the conjunction "and" in various contexts

No Comments

No Known Activations