INDEX

Explanations

words related to providing evidence or support for a claim

oai_token-act-pair · gpt-3.5-turbo

terms related to substantiation and refutation in arguments

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

 Mistress

-0.70

 Drawn

-0.69

killer

-0.67

eries

-0.67

fork

-0.62

izabeth

-0.62

helm

-0.61

>>>>>>>>

-0.61

crow

-0.61

nesday

-0.61

POSITIVE LOGITS

acles

0.99

ivity

0.96

acle

0.93

oret

0.92

ctr

0.88

iveness

0.87

race

0.87

iating

0.87

raint

0.84

iation

0.84

Activations Density 0.016%

words related to providing evidence or support for a claim

terms related to substantiation and refutation in arguments

No Comments

No Known Activations