INDEX

Explanations

words related to investigation or examination

oai_token-act-pair · gpt-3.5-turbo

significant terms related to political discussions and controversies

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 pecul

-0.70

++++++++++++++++

-0.67

_.

-0.66

 triv

-0.65

)\

-0.64

↵Âł

-0.63

////////////////////////////////

-0.63

Secondly

-0.61

theless

-0.61

)).

-0.61

POSITIVE LOGITS

Enlarge

0.77

oward

0.72

 amid

0.71

 toward

0.69

 against

0.67

 regarding

0.66

to

0.62

 Thursday

0.61

 behind

0.60

 Tuesday

0.59

Activations Density 0.643%

words related to investigation or examination

significant terms related to political discussions and controversies

No Comments

No Known Activations