INDEX

Explanations

mentions of a specific term, "Sa", which is not related to a specific concept in this context

oai_token-act-pair · gpt-3.5-turbo

mentions of the name "Sa" 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 @ 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

papers

-0.79

ãĥ¼ãĥĨãĤ£

-0.78

tics

-0.78

theless

-0.72

lessly

-0.72

breaks

-0.72

mercial

-0.72

å§«

-0.70

 Turing

-0.66

tyard

-0.65

POSITIVE LOGITS

igon

0.98

uten

0.98

adish

0.98

iva

0.97

Ga

0.92

uth

0.89

vers

0.89

Sa

0.89

ivas

0.88

pling

0.88

Activations Density 0.010%

mentions of a specific term, "Sa", which is not related to a specific concept in this context

mentions of the name "Sa" in various contexts

No Comments

No Known Activations