INDEX

Explanations

verbs related to observation and decision-making

oai_token-act-pair · gpt-3.5-turbo

instances of the word "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 @ 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

ENCY

-0.76

å¤©

-0.70

BAT

-0.69

Shut

-0.69

!--

-0.64

FOR

-0.63

endo

-0.63

LOS

-0.63

SI

-0.62

apo

-0.62

POSITIVE LOGITS

 evaluate

0.97

 analyse

0.96

 analyze

0.94

 evaluating

0.92

 discern

0.91

 determine

0.88

 concluded

0.85

 assessing

0.85

 analyzing

0.85

 analysed

0.84

Activations Density 0.381%

verbs related to observation and decision-making

instances of the word "and" in various contexts

No Comments

No Known Activations