INDEX

Explanations

phrases related to goals, achievements, promises, and benefits

oai_token-act-pair · gpt-3.5-turbo

terms related to crime and consequences

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

stable

-0.78

arling

-0.77

ateurs

-0.74

lip

-0.74

ovy

-0.74

annis

-0.72

untled

-0.68

orthy

-0.67

 PACs

-0.66

aples

-0.65

POSITIVE LOGITS

 refrain

0.73

 imaginable

0.72

ãĥĨãĤ£

0.72

phrase

0.69

 spree

0.68

lessly

0.66

IENCE

0.65

lessness

0.64

 whereby

0.64

ously

0.63

Activations Density 0.546%

phrases related to goals, achievements, promises, and benefits

terms related to crime and consequences

No Comments

No Known Activations