INDEX

Explanations

the word "never" followed by a verb, often indicating a negative action or state

oai_token-act-pair · gpt-3.5-turbo

repeated instances of the word "never" 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 @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

Notes

-0.63

HP

-0.60

uing

-0.59

assing

-0.57

 Towards

-0.56

 Shutdown

-0.55

States

-0.54

thening

-0.54

 Journals

-0.54

VL

-0.54

POSITIVE LOGITS

been

1.24

 been

1.22

theless

1.02

 Been

0.99

 tasted

0.90

 gotten

0.87

 undergone

0.85

 existed

0.83

 seen

0.80

 heard

0.79

Activations Density 0.045%

the word "never" followed by a verb, often indicating a negative action or state

repeated instances of the word "never" in various contexts

No Comments

No Known Activations